With the tightening of the federal budget, it is becoming increasingly more difficult to secure funding for research. We are therefore very happy to report that two projects by Michigan Linguistics faculty have recently been funded by the National Science Foundation. We will report in a future edition of the news blog about Carmel O’Shannessy’s project on Warlpiri. This report will focus on a joint project by Pam Beddor and Andries Coetzee.

Pam and Andries’s project is called The time course of speech perception and production in individual language users. Their project capitalizes on prior experience in the Department with the perception and processing of coarticulated speech, but it also pushes into new, still unexplored territories. They will explore the relation between patterns of production and perception of coarticulated speech in individual language users. Understanding this relation better will inform theories of sound change. In addition to studying several coarticulated structures in English, Pam and Andries will investigate the perception and production of anticipatory nasalization in Afrikaans. Not only will this extend the study of speech production and perception to an under-described language, but due to the complex social structure of the Afrikaans speech community this will also afford them the opportunity to explore the role of sociolinguistic factors in the perception of coarticulated speech.

Pam and Andries will use various techniques in their project, ranging from acoustic analysis of speech, to the study of articulation through ultrasound imaging and aerodynamic measures, to the study of speech processing through eye-tracking. The award will enable them to purchase new equipment for the Phonetics Laboratory, including a new state-of-the-art eye-tracker. Additionally, the award also contains funding for a post-doctoral researcher (a search for a post-doc is currently underway).

The abstract of the award, as published on the NSF website, is given below.


The time course of speech perception and production in individual language users

Clarifying the highly complex dynamic processes that underlie the production and perception of speech is critical both to our understanding of human language processing and to improvements in technologies for computer-mediated communication. Speakers must fluently coordinate the actions of multiple vocal organs (such as the lips, tongue, and jaw) to produce the overlapping movements necessary for smooth and rapid speech. In deriving meaning from the resulting acoustic signal, listeners must attend to the changing acoustic properties as they evolve over time.

With the support of the National Science Foundation, Dr. Beddor and Dr. Coetzee are studying the relation between the dynamics of spoken language production and perception. Previous research has found that speakers differ from each other in their precise patterns of articulatory coordination, and that listeners differ in their sensitivity to those variable coarticulated patterns in deciding what a speaker has said. This project investigates whether there is a link between a listener's use of coarticulatory information in perception and that language user's own coarticulated productions. 

Guided by the hypothesis that the relation between a language user's perception and production is mediated by social, cognitive, and other factors, a series of experiments tests both socially neutral and socially indexed patterns of articulatory coordination. These experiments also test language users' perceptual awareness of, and articulatory accommodation to, new patterns of coordination. Listeners' real-time processing of the acoustic signal is monitored using eye-tracking methods; the time course of production is assessed via airflow, ultrasound imaging, and acoustic analysis. The project will gather data on the production and perception of America English and Afrikaans. The Afrikaans work will broaden empirical coverage of a phonetically understudied language and will extend collaboration with South African researchers. An overarching goal of studying the production-perception relation in individual language users is to understand how linguistic structures are represented in the human mind. Elucidating this relation in its social context should also contribute to an understanding of how individual differences in these structures might serve as a source of new sound patterns that spread through a speech community.